CN116644848A - Multi-source data-based medium-long term area wind power prediction method - Google Patents
Multi-source data-based medium-long term area wind power prediction method Download PDFInfo
- Publication number
- CN116644848A CN116644848A CN202310590958.XA CN202310590958A CN116644848A CN 116644848 A CN116644848 A CN 116644848A CN 202310590958 A CN202310590958 A CN 202310590958A CN 116644848 A CN116644848 A CN 116644848A
- Authority
- CN
- China
- Prior art keywords
- data
- wind power
- prediction
- power
- long term
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000013135 deep learning Methods 0.000 claims abstract description 12
- 238000004364 calculation method Methods 0.000 claims abstract description 5
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000010606 normalization Methods 0.000 claims description 17
- 238000013527 convolutional neural network Methods 0.000 claims description 11
- 210000002569 neuron Anatomy 0.000 claims description 10
- 230000005855 radiation Effects 0.000 claims description 10
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000007774 longterm Effects 0.000 claims description 8
- 230000002779 inactivation Effects 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 7
- 238000004140 cleaning Methods 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 5
- 238000001556 precipitation Methods 0.000 claims description 5
- 238000012360 testing method Methods 0.000 claims description 4
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 238000010200 validation analysis Methods 0.000 claims 1
- 238000011160 research Methods 0.000 abstract description 5
- 230000002159 abnormal effect Effects 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012795 verification Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/10—Pre-processing; Data cleansing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/251—Fusion techniques of input or preprocessed data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/0985—Hyperparameter optimisation; Meta-learning; Learning-to-learn
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/004—Generation forecast, e.g. methods or systems for forecasting future energy generation
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/10—Power transmission or distribution systems management focussing at grid-level, e.g. load flow analysis, node profile computation, meshed network optimisation, active network management or spinning reserve management
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2203/00—Indexing scheme relating to details of circuit arrangements for AC mains or AC distribution networks
- H02J2203/20—Simulating, e g planning, reliability check, modelling or computer assisted design [CAD]
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J2300/00—Systems for supplying or distributing electric power characterised by decentralized, dispersed, or local generation
- H02J2300/20—The dispersed energy generation being of renewable origin
- H02J2300/28—The renewable source being wind energy
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Business, Economics & Management (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Economics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Strategic Management (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Evolutionary Biology (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Human Resources & Organizations (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Primary Health Care (AREA)
- Water Supply & Treatment (AREA)
- Public Health (AREA)
- Quality & Reliability (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Power Engineering (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The application provides a multi-source data-based medium-long term regional wind power prediction method, which is based on deep learning mixed mode of multi-source dataA mold, comprising: s is S 1 Collecting numerical weather forecast data required by a prediction model; s is S 2 Preprocessing the collected numerical weather forecast data, and performing splicing arrangement; s is S 3 Calculating the predicted power of the similar day according to the numerical weather forecast data and the historical power data of the matched time, and splicing the predicted power and the time influence factor into a vector; s is S 4 Loading a pre-trained multi-network combined prediction model, and performing the step S 2 Processed data and step S 3 The processed data are spliced and input into a prediction model; s is S 5 The prediction model returns a prediction result of wind power prediction through forward calculation; s is S 6 And converting the plurality of values of the predicted result into daily average power. The method solves the problem of short wind power prediction time, and can provide key prediction information for middle-long term power contract transaction strategy research and judgment.
Description
Technical Field
The application relates to the field of wind power prediction methods, in particular to a medium-long term regional wind power prediction method based on multi-source data.
Background
In the prior art, with the continuous development of wind power related technologies and power market reform in recent years, the wind power generation industry in China has been developed at a high speed, and the medium-and long-term power market has been opened. However, most of wind power prediction products in the market at present are short-term or ultra-short-term wind power predictions, and research on mid-long-term wind power predictions is less. Predicting the long-term wind power in the area is helpful for providing references for planning and decomposing wind power trade in the wind power medium-term and long-term power trade.
However, there are certain difficulties in mid-to-long term wind power predictions compared to short-term or ultra-short term wind power predictions. The available years are limited due to the historical power of wind power and the influence data related to the predicted height of the wind power, and the installed capacity of the wind power which is increased in the years is large and irregular due to the rapid development of the wind power. In mid-to-long term wind power prediction, mid-to-long term climate pattern data is an important factor available to affect power prediction fluctuations.
In view of the above, the present inventors devised a method for predicting wind power in a medium-long term region based on multi-source data, so as to overcome the above-mentioned technical problems.
Disclosure of Invention
The application aims to overcome the defect that a medium-long term wind power prediction product is lacking in the prior art, and provides a medium-long term region wind power prediction method based on multi-source data.
The application solves the technical problems by the following technical proposal:
the medium-long term region wind power prediction method based on the multi-source data is characterized by comprising the following steps of:
S 1 collecting numerical weather forecast data required by a prediction model;
S 2 preprocessing the collected numerical weather forecast data, converting two-dimensional image data into three-dimensional image data for superposition, and splicing and arranging;
S 3 calculating the predicted power of the similar day according to the numerical weather forecast data and the historical power data of the matched time, and splicing the predicted power and the time influence factor into a vector;
S 4 loading a pre-trained multi-network combined prediction model, and carrying out the step S 2 Processed data and step S 3 The processed data are spliced and then input into the prediction model;
S 5 the prediction model returns a prediction result of wind power prediction through forward calculation;
S 6 and converting the plurality of values of the prediction result into daily average power.
According to one embodiment of the application, the step S 1 Comprises the following steps: historical data of relevant influence factors influencing wind power prediction of an electric power market and actual clear power data of a power grid are collected.
According to one embodiment of the application, the historical data of the relevant influence factors includes wind speed and direction, surface temperature, near-ground air pressure, ground precipitation rate, near-ground air specific humidity, sea level air pressure, ground down short wave radiation, ground down long wave radiation, ground sea level air pressure and total cloud cover of different height layers.
According to one embodiment of the application, the step S 2 Comprises the following steps: and carrying out data cleaning, data cutting, splicing and normalization on the screened factors and the actual wind power prediction data.
According to one embodiment of the application, the data cleaning comprises filling or rejecting missing values and abnormal values of data; the data cutting corresponds the data to the area range needing to be predicted; the data splicing performs spatial superposition on the required factor variables;
the data normalization compresses the data to be normalized between [0,1], and the formula is as follows:
wherein ,is the normalized input variable; x is x max Is the maximum value of the variable x; x is x min Is the minimum value of the variable x.
According to one embodiment of the application, the step S 3 Comprises the following steps: and (3) the relevant influence factors are time-to-time corresponding to the actual power, the actual power on similar days is searched through weather conditions and is used as predicted power, and meanwhile, the predicted power is combined with the time dimension factors to be spliced.
According to one embodiment of the application, the step S 3 Also comprises: grouping the dataThe data after the integration is processed into a time sequence corresponding to the date of the input sample formed by the multidimensional characteristic factors and the output sample with the wind power predicted as the true value.
According to one embodiment of the application, the sequence of the input samples is x= [ [ a ] (1) ,a (2) ,a (3) ,......,a (m) ],[b (1) ,b (2) ,b (3) ,......,b (m) ]]；
wherein ,a(m) B, combining all historical time data of each variable of weather forecast data corresponding to the mth day into a time sequence sample (m) A time series sample formed by combining the predicted power and the time influence factor of the similar day corresponding to the time of the mth day;
the wind power prediction sequence of the corresponding output sample, namely the corresponding history for a certain day is Y= [ Y ] 1 ,y 2 ,y 3 ,......,y m ]，y m And predicting the historical wind power corresponding to the m-th day.
According to one embodiment of the application, the step S 4 Comprises the following steps: dividing the processed data set according to 7:1.5:1.5, wherein the training set is used for model fitting, the verification set is used for model hyper-parameter adjustment, and the test set is used for checking whether the trained model has generalization capability.
According to one embodiment of the application, the step S 4 Also comprises: determining a network structure of deep learning, adjusting network parameters, training a network by adjusting the number of network layers, the number of neurons, batch normalization and random inactivation until the maximum iteration number or the convergence of the network learning rate is reached;
the mean square error is selected as a loss function, and the average absolute error is selected as an evaluation precision:
where MSE represents the mean square error and MAE represents the mean absolute error.
According to one embodiment of the application, the prediction model is a deep learning hybrid model, comprising a convolutional neural network and a deep neural network;
the input layer of the convolutional neural network is data spliced by various meteorological factors in each time of the day, and the convolutional neural network is provided with two-dimensional convolutional layers;
the model structure of the deep neural network comprises an input layer, an activation layer, a middle layer and an output layer, and meanwhile, a mode of combining batch normalization and random inactivation items is selected to prevent the model from overfitting in advance, so that the robustness of the model is enhanced.
The application has the positive progress effects that:
according to the method for predicting the wind power in the medium-long-term area based on the multi-source data, weather forecast mode data of meteorological values are combined, the average daily power is used as a forecast object based on a deep learning method, the wind power prediction for 60 days in the future is provided, and critical prediction information can be provided for medium-long-term wind power trade.
Compared with wind power prediction products on the market, the wind power prediction method for the medium-long term area solves the problem of short wind power prediction time, and can provide key prediction information for the research and judgment of medium-long term power contract trading strategies.
Drawings
The above and other features, properties and advantages of the present application will become more apparent from the following description of embodiments taken in conjunction with the accompanying drawings in which like reference characters designate like features throughout the drawings, and in which:
FIG. 1 is a flow chart of a method for predicting wind power in a medium-long term area based on multi-source data.
FIG. 2 is a flow chart of a prediction model in the mid-long term area wind power prediction method based on multi-source data.
Fig. 3 is a schematic diagram of a region wind power prediction deep learning hybrid model in the multi-source data-based medium-long term region wind power prediction method of the present application.
Detailed Description
In order to make the above objects, features and advantages of the present application more comprehensible, embodiments accompanied with figures are described in detail below.
Embodiments of the present application will now be described in detail with reference to the accompanying drawings. Reference will now be made in detail to the preferred embodiments of the present application, examples of which are illustrated in the accompanying drawings. Wherever possible, the same reference numbers will be used throughout the drawings to refer to the same or like parts.
Furthermore, although terms used in the present application are selected from publicly known and commonly used terms, some terms mentioned in the present specification may be selected by the applicant at his or her discretion, the detailed meanings of which are described in relevant parts of the description herein.
Furthermore, it is required that the present application is understood, not simply by the actual terms used but by the meaning of each term lying within.
As shown in fig. 1, the application discloses a multi-source data-based medium-long term regional wind power prediction method, which is based on a multi-source data deep learning hybrid model and comprises the following steps:
step S 1 And collecting numerical weather forecast data, namely weather factor data, required by the prediction model.
The numerical weather forecast data mainly comprises wind speed and direction, surface temperature, near-ground air pressure, ground precipitation rate, near-ground air specific humidity, sea level air pressure, ground downward short wave radiation, ground downward long wave radiation, ground sea level air pressure, total cloud amount and other relevant factors of different height layers.
Wherein, preferably, the step S 1 Comprises the following steps: historical data of relevant influence factors influencing wind power prediction of an electric power market and actual clear power data of a power grid are collected.
The related influence factor data comprise wind speed and direction, surface temperature, near-ground air pressure, ground precipitation rate, near-ground air specific humidity, sea level air pressure, ground downward short wave radiation, ground downward long wave radiation, ground sea level air pressure, total cloud amount and the like of different height layers.
Step S 2 Preprocessing the collected numerical weather forecast data, converting the two-dimensional image data into three-dimensional image data for superposition, and performing splicing arrangement.
Preferably, the step S 2 Comprises the following steps: and carrying out data cleaning, data cutting, splicing and normalization on the screened factors and the actual wind power prediction data.
And the data cleaning comprises filling or removing missing values and abnormal values of the data. The data clipping corresponds the data to the range of regions that need to be predicted. And the data splicing performs spatial superposition on the required factor variables.
The data normalization compresses the data to be normalized between [0,1], and the formula is as follows:
wherein ,is the normalized input variable; x is x max Is the maximum value of the variable x; x is x min Is the minimum value of the variable x.
Step S 3 And calculating the predicted power of the similar day according to the numerical weather forecast data (namely the meteorological factor data) and the historical power data of the matched time, and splicing the predicted power and the time influence factor into a vector. For example, the time influence factor is the time of day of the year, and the time of day position is spliced into a vector.
And (3) the factors are corresponding to the actual power time, and the actual power on the similar day is searched through weather conditions and used as predicted power. And simultaneously, splicing is carried out by combining the time dimension factors.
In addition, the step S 3 Also comprises: and processing the data after data normalization into a time sequence corresponding to the date of an input sample formed by the multidimensional characteristic factors and an output sample with wind power predicted to be a true value.
The time interval of the wind power prediction data is 15 minutes, 96 values are obtained in one day, and the resolution of the numerical weather prediction data of the mode output is difficult to achieve one-to-one prediction, so that the mode prediction data of all times in one day are used for predicting the wind power of n (n is more than or equal to 1 and less than or equal to 96 and n is an integer) points in one day.
The sequence of the input samples is X= [ [ a ] (1) ,a (2) ,a (3) ,......,a (m) ],[b (1) ,b (2) ,b (3) ,......,b (m) ]]；
wherein ,a(m) B, combining all historical time data of each variable of weather forecast data corresponding to the mth day into a time sequence sample (m) And predicting power and time influence factors for the similar days corresponding to the time of the mth day to form a time series sample. The formula is as follows, a (m) ＝[a 1 ,a 2 ,a 3 ,......,a t ]，b (m) ＝[b 1 ,b 2 ,b 3 ,......,b t ]T is the corresponding time. The wind power prediction sequence of the corresponding output sample, namely the corresponding history for a certain day is Y= [ Y ] 1 ,y 2 ,y 3 ,......,y m ]，y m And predicting the historical wind power corresponding to the m-th day.
Step S 4 Loading a pre-trained multi-network combined prediction model, and carrying out the step S 2 Processed data and step S 3 And splicing the processed data, and then inputting the spliced data into the prediction model.
Preferably, the step S 4 Comprises the following steps: dividing the processed data set according to 7:1.5:1.5, wherein the training set is used for model fitting, the verification set is used for model hyper-parameter adjustment, and the test set is used for checking whether the trained model has generalization capability.
Next, the step S 4 Also comprises: determining a deep learning network structure, adjusting network parameters, training the network by adjusting the number of network layers, the number of neurons, batch normalization BatchNormalization (BN) and random inactivation Dropout (DP) until the maximum iteration number or the network learning rate converges.
Mean square error (Mean Square Error, MSE) was chosen as the loss function, mean absolute error (Mean Absolute Error, MAE) as the evaluation accuracy:
where MSE represents the mean square error and MAE represents the mean absolute error.
The prediction model is a deep learning hybrid model and comprises a convolutional neural network and a deep neural network.
The specific structure of the convolutional neural network is as follows: the input layer of the convolutional neural network is data spliced by various meteorological factors in each time of the day, and has a certain length, width and depth. Second, the convolutional neural network has two-dimensional convolutional layers.
The specific calculation result of the convolution has the specific formula:
wherein: l is the number of layers of the convolution layer, i and j correspond to the length and width of the image data respectively,is the convolution kernel of layer I, +.>As bias term, F j Is an input feature.
The filter sizes of the two convolution layers are set to be 3 multiplied by 3, wherein 16 filters are used for the first convolution layer, a sampling layer with a pooling layer size of 2 multiplied by 2 is arranged behind the first convolution layer as output, 32 filters are used for the second convolution layer, batch normalization (batch normalization) items are arranged behind the first convolution layer, an activation function is set to be Relu, and nonlinear conversion is completed on characteristic information, and the specific formula is as follows: relu (x) =max (0, x). Finally, 2 full-connection layers are arranged, the multidimensional features are paved and unfolded, and the number of neurons is 1024 and 512 respectively.
The model structure of the deep neural network comprises an input layer, an activation layer, a middle layer and an output layer, and meanwhile, a mode of combining batch normalization BatchNormalization (BN) and random inactivation Dropout (DP) items is selected to prevent the model from exceeding fitting phenomenon in advance, so that the robustness of the model is enhanced.
The input of the deep neural network is the data spliced in the step 3, and three layers of neural networks are arranged, wherein the number of neurons can be preferably 16, 32 and 64 respectively. Of course, the number of neurons can be fine-tuned according to the application in different area scenarios, and the above number is only used as an example and is not limiting on the number of neurons. The first layer network is followed by a batch normalization term, with each layer activation function set to Relu.
The final model was the combination of the outputs of the two models, set Dropout term to 0.5, and add 2 full connection layers, with neuron numbers 512, 256. Of course, the number of neurons can be fine-tuned according to the application in different area scenarios, and the above number is only used as an example and is not limiting on the number of neurons. The Dropout (random inactivation) terms are all 0.3, and the final output layer is n, corresponding to the wind power prediction at the required moment. And (3) performing iterative optimization by using an Adam algorithm until the numerical value of the loss function converges or reaches the maximum iterative times, and completing model training.
Step S 5 And the prediction model returns a prediction result of wind power prediction through forward calculation.
Step S 6 And converting the plurality of values of the prediction result into daily average power.
Step S 5 And step S 6 For practical use, by combining step S 3 Data input step S of the process 4 In the model of (2), a wind power predicted value is obtained, and finally, a daily average is taken as a final wind power predicted value.
The application relates to a multi-source data-based medium-long term regional wind power prediction method, which comprehensively considers factors influencing wind power prediction such as wind speed and direction, surface temperature, near-ground air pressure, ground precipitation rate, near-ground air specific humidity, sea level air pressure, ground downward short wave radiation, ground downward long wave radiation, ground sea level air pressure, total cloud quantity and the like of different height layers and power values of similar days, and establishes a wind power prediction model for predicting the future 60 days by combining a deep learning convolutional neural network and a deep neural network combined model.
In summary, the method for predicting wind power in the middle-long-term area based on the multi-source data combines weather value weather forecast mode data, takes daily average power as a forecast object based on a deep learning method, provides wind power prediction for 60 days in the future, and can provide critical prediction information for middle-long-term wind power trade.
Compared with wind power prediction products on the market, the wind power prediction method for the medium-long term area solves the problem of short wind power prediction time, and can provide key prediction information for the research and judgment of medium-long term power contract trading strategies. The medium-long term regional wind power prediction method solves the problem of current lack of medium-long term wind power prediction products, and provides key prediction information for the research and judgment of medium-long term power contract trading strategies.
The application uses specific words to describe embodiments of the application. Reference to "one embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the application. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the application may be combined as suitable.
While specific embodiments of the application have been described above, it will be appreciated by those skilled in the art that these are by way of example only, and the scope of the application is defined by the appended claims. Various changes and modifications to these embodiments may be made by those skilled in the art without departing from the principles and spirit of the application, but such changes and modifications fall within the scope of the application.
Claims (11)
1. The medium-long term region wind power prediction method based on the multi-source data is characterized by comprising the following steps of:
S 1 collecting numerical weather forecast data required by a prediction model;
S 2 preprocessing the collected numerical weather forecast data, converting two-dimensional image data into three-dimensional image data for superposition, and splicing and arranging;
S 3 calculating the predicted power of the similar day according to the numerical weather forecast data and the historical power data of the matched time, and splicing the predicted power and the time influence factor into a vector;
S 4 loading a pre-trained multi-network combined prediction model, and carrying out the step S 2 Processed data and step S 3 The processed data are spliced and then input into the prediction model;
S 5 the prediction model returns a prediction result of wind power prediction through forward calculation;
S 6 and converting the plurality of values of the prediction result into daily average power.
2. The multi-source data based mid-to-long term zone wind power prediction of claim 1The method is characterized in that the step S 1 Comprises the following steps: historical data of relevant influence factors influencing wind power prediction of an electric power market and actual clear power data of a power grid are collected.
3. The method for predicting wind power in a medium-long term area based on multi-source data according to claim 2, wherein the historical data of the relevant influence factors comprises wind speed and direction of different height layers, surface temperature, near-ground air pressure, ground precipitation rate, near-ground air specific humidity, sea level air pressure, ground down short wave radiation, ground down long wave radiation, ground sea level air pressure and total cloud amount.
4. The method for predicting wind power in medium and long term areas based on multi-source data according to claim 2, wherein said step S 2 Comprises the following steps: and carrying out data cleaning, data cutting, splicing and normalization on the screened factors and the actual wind power prediction data.
5. The multi-source data-based medium-long term regional wind power prediction method of claim 4, wherein the data cleaning comprises filling or rejecting missing values and outliers of the data; the data cutting corresponds the data to the area range needing to be predicted; the data splicing performs spatial superposition on the required factor variables;
the data normalization compresses the data to be normalized between [0,1], and the formula is as follows:
wherein ,is the normalized input variable; x is x max Is the maximum value of the variable x; x is x min Is the minimum value of the variable x.
6. The method for predicting wind power in medium and long term areas based on multi-source data as set forth in claim 5, wherein said step S 3 Comprises the following steps: and (3) the relevant influence factors are time-to-time corresponding to the actual power, the actual power on similar days is searched through weather conditions and is used as predicted power, and meanwhile, the predicted power is combined with the time dimension factors to be spliced.
7. The method for predicting wind power in medium and long term areas based on multi-source data as set forth in claim 6, wherein said step S 3 Also comprises: and processing the data after data normalization into a time sequence corresponding to the date of an input sample formed by the multidimensional characteristic factors and an output sample with wind power predicted to be a true value.
8. The method for predicting wind power in mid-long term region based on multi-source data of claim 7, wherein the sequence of input samples is x= [ [ a ] (1) ,a (2) ,a (3) ,......,a (m) ],[b (1) ,b (2) ,b (3) ,......,b (m) ]]；
wherein ,a(m) B, combining all historical time data of each variable of weather forecast data corresponding to the mth day into a time sequence sample (m) A time series sample formed by combining the predicted power and the time influence factor of the similar day corresponding to the time of the mth day;
the wind power prediction sequence of the corresponding output sample, namely the corresponding history for a certain day is Y= [ Y ] 1 ,y 2 ,y 3 ,......,y m ]，y m And predicting the historical wind power corresponding to the m-th day.
9. The method for predicting wind power in medium and long term areas based on multi-source data as set forth in claim 1, wherein said step S 4 Comprises the following steps: dividing the processed data set according to 7:1.5:1.5, the training set is used for model fitting, the validation set is used for model fitting, and the test set is divided intoAnd (3) model hyper-parameter adjustment, wherein a test set is used for checking whether the trained model has generalization capability.
10. The method for predicting wind power in medium and long term areas based on multi-source data as set forth in claim 9, wherein said step S 4 Also comprises: determining a network structure of deep learning, adjusting network parameters, training a network by adjusting the number of network layers, the number of neurons, batch normalization and random inactivation until the maximum iteration number or the convergence of the network learning rate is reached;
the mean square error is selected as a loss function, and the average absolute error is selected as an evaluation precision:
where MSE represents the mean square error and MAE represents the mean absolute error.
11. The multi-source data based medium-long term region wind power prediction method of claim 10, wherein the prediction model is a deep learning hybrid model comprising a convolutional neural network and a deep neural network;
the input layer of the convolutional neural network is data spliced by various meteorological factors in each time of the day, and the convolutional neural network is provided with two-dimensional convolutional layers;
the model structure of the deep neural network comprises an input layer, an activation layer, a middle layer and an output layer, and meanwhile, a mode of combining batch normalization and random inactivation items is selected to prevent the model from overfitting in advance, so that the robustness of the model is enhanced.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310590958.XA CN116644848A (en) | 2023-05-24 | 2023-05-24 | Multi-source data-based medium-long term area wind power prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310590958.XA CN116644848A (en) | 2023-05-24 | 2023-05-24 | Multi-source data-based medium-long term area wind power prediction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116644848A true CN116644848A (en) | 2023-08-25 |
Family
ID=87624104
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310590958.XA Pending CN116644848A (en) | 2023-05-24 | 2023-05-24 | Multi-source data-based medium-long term area wind power prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116644848A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117060407A (en) * | 2023-10-12 | 2023-11-14 | 国网湖北省电力有限公司经济技术研究院 | Wind power cluster power prediction method and system based on similar day division |
| CN117290810A (en) * | 2023-11-27 | 2023-12-26 | 南京气象科技创新研究院 | Short-time strong precipitation probability prediction fusion method based on cyclic convolutional neural network |
-
2023
- 2023-05-24 CN CN202310590958.XA patent/CN116644848A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117060407A (en) * | 2023-10-12 | 2023-11-14 | 国网湖北省电力有限公司经济技术研究院 | Wind power cluster power prediction method and system based on similar day division |
| CN117060407B (en) * | 2023-10-12 | 2023-12-22 | 国网湖北省电力有限公司经济技术研究院 | Wind power cluster power prediction method and system based on similar day division |
| CN117290810A (en) * | 2023-11-27 | 2023-12-26 | 南京气象科技创新研究院 | Short-time strong precipitation probability prediction fusion method based on cyclic convolutional neural network |
| CN117290810B (en) * | 2023-11-27 | 2024-02-02 | 南京气象科技创新研究院 | Short-time strong precipitation probability prediction fusion method based on cyclic convolutional neural network |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111310889B (en) | Evaporation waveguide profile estimation method based on deep neural network | |
| Fan et al. | Empirical and machine learning models for predicting daily global solar radiation from sunshine duration: A review and case study in China | |
| Zhou et al. | A review on global solar radiation prediction with machine learning models in a comprehensive perspective | |
| Haq et al. | A new hybrid model for short-term electricity load forecasting | |
| CN116644848A (en) | Multi-source data-based medium-long term area wind power prediction method | |
| CN111126704B (en) | Multi-region precipitation prediction model construction method based on multi-graph convolution and memory network | |
| Raza et al. | An ensemble framework for day-ahead forecast of PV output power in smart grids | |
| Quej et al. | ANFIS, SVM and ANN soft-computing techniques to estimate daily global solar radiation in a warm sub-humid environment | |
| Lipu et al. | Artificial intelligence based hybrid forecasting approaches for wind power generation: Progress, challenges and prospects | |
| Ramirez et al. | Artificial neural network technique for rainfall forecasting applied to the Sao Paulo region | |
| Hocaoğlu et al. | Hourly solar radiation forecasting using optimal coefficient 2-D linear filters and feed-forward neural networks | |
| CN112288164B (en) | Wind power combined prediction method considering spatial correlation and correcting numerical weather forecast | |
| CN109523013B (en) | Air particulate matter pollution degree estimation method based on shallow convolutional neural network | |
| CN111160520A (en) | BP neural network wind speed prediction method based on genetic algorithm optimization | |
| CN112287294B (en) | Space-time bidirectional soil water content interpolation method based on deep learning | |
| CN111695731A (en) | Load prediction method, system and equipment based on multi-source data and hybrid neural network | |
| CN109143408B (en) | Dynamic region combined short-time rainfall forecasting method based on MLP | |
| CN113705877A (en) | Real-time monthly runoff forecasting method based on deep learning model | |
| CN115049024B (en) | Training method and device of wind speed prediction model, electronic equipment and storage medium | |
| CN114298134A (en) | Wind power prediction method and device and electronic equipment | |
| CN105844334B (en) | A kind of temperature interpolation method based on radial base neural net | |
| CN115019510A (en) | Traffic data restoration method based on dynamic self-adaptive generation countermeasure network | |
| CN111597751A (en) | Crude oil film absolute thickness inversion method based on self-expansion depth confidence network | |
| Rana et al. | A data-driven approach for forecasting state level aggregated solar photovoltaic power production | |
| CN115186923A (en) | Photovoltaic power generation power prediction method and device and electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |